Everything you need to know about 2025, from an industry so fast it needs its own theory of relativity. Daniel Purnell of SmartAssets takes stock with the expert help of his colleagues.
“Out with the old, in with the new” isn’t a motto for the beginning of each year in martech – it’s every Monday morning. Before the coffee has been poured and you open your first email, there’s something new.
A new technology to test. A client who wants you to innovate a new solution. A need to pivot. With that in mind, here’s how to better navigate tech in 2026.
The early tech-adopter gets the worm, or at least a meaningful head start
Early 2025 was characterized by a proliferation of AI, but not everybody was on board. As data quality and AI integration became a top priority, the majority of organizations were already falling behind in a classic scramble of ‘too little too late.’
Doug Guttenberg, EVP of integrated studios at Doner, notes that 2025 was the best time to debut ‘AI operationalization,’ however, most teams: “underestimated how quickly the gap would open between those who experimented early, and those who waited.” The aim wasn’t to be perfect with your first build, agent, or tool, but to actually make leeway.
By the second quarter, everybody was obsessing over agentic AI. “Agents were sold as autonomous experts,” says Pete Steiner, group creative director at Code and Theory. But what they didn’t realize is that: “AI agents are just flexible automation systems that still need structured data, clear workflows, and thoughtful setup. They’re more capable than traditional automation, but they’re still automation.”
“Their real value lies in the unglamorous work every organization struggles with,” he adds. “We never saw the agent that can generate a full campaign from a brief, and I doubt we ever will. Agents matter, and their impact will keep growing, yet we need more grounded expectations.”
SmartAssets’ CEO Lindsay Hong agrees, warning that we often become too focused on the next big shiny thing, yet: “it doesn’t always live up to expectations.” This was the case with in-flight optimization, she says. “Marketers got far too excited about in-flight testing this year, but the premise of these solutions is built on funding failures.”
Hong is referring to the number of assets that don’t perform, yet advertisers have had to spend heavily on these to find out which ones see an uplift. “Using historical data to predict the success of assets preflight prevents wastage and offers much better strategic insight,” she explains.
Sometimes the outlandish can quickly become the new norm
The industry was quick to pick holes in generative AI’s output. After a short-lived excitement of its capabilities, everyone jumped on the bandwagon of “look at the seven fingers” or “it can’t even handle text,” remembers Hong. But since then, digital mountains have been moved.
“Earlier this year, there was no solution for applying text overlay graphics to a photorealistic image,” she says. “Fast-forward just a few months, and we can create media-ready assets with logos, text, and hyper-realistic backgrounds. I think we massively underestimated how quickly that would come about.”
Want to go deeper? Ask The Drum
Guttenberg draws parallels between the quick dismissal of AI and vibe coding. “At the start of the year, it felt like a fringe idea, but by Q4 it became a default way to get something built. It opened the door for non-technical thinkers to turn an idea into a working version in a single sitting,” he explains.
The big takeaway here is to not overlook the potential of new technologies: developers, marketers, and many in between have quickly changed their tune in the last 12 months. “Every quarter, something groundbreaking happens,” adds Guttenberg. So, what may seem like an outlandish idea to start with may end up the new reality before you know it.
Did gen AI overpromise in creative production?
The expectations for gen-AI were grandiose. After all, the AI market is set to exceed $740bn in the next decade. But whether or not it’s living up to its creative production expectations is the big question.
Guttenberg believes that although generative AI has made “good progress,” when it comes down to churning out assets and variations, we’re still “missing the mark” for a few reasons:
1) Data isn’t connected enough: from first and third-party data to audience strategy, media targeting, and generative-AI capabilities haven’t been properly built, creating ambiguity.
2) Critical mass of quality: systems aren’t ready to operate without oversight just yet. But, should recent history be anything to go by, this is likely to change in the next few quarters.
3) Comfort levels: marketing teams need to triple-check accuracy and feel in control. However, as data and quality improve (to the above points), overall outputs will become more consistent.
The feeling of missed opps echoes through the creative industry. “High-volume variation has existed for years through templates and automation,” highlights Steiner. “Whereas gen AI adds flexibility and creative range, the core challenge remains: identifying why a variation exists and who it’s for, then proving whether it works.”
“AI personas are improving the ‘why,’ and governance and optimization tools such as SmartAssets are improving quality control and media effectiveness. The next wave will be campaigns that finally evolve in real time, especially in the mid and lower funnel where variation and scale matter most.”
Another challenge with gen AI is anxiety about not catching the potential errors it could generate, mixed with the FOMO of being left behind. This is what’s really stifling creativity. “If we can provide more reassurances that we can catch non-compliant assets, then more energy, time, and focus can go into the creative ideas that actually drive engagement,” says Hong.
The most promising technologies to look out for in 2026
If the dizzying array of technologies to choose from wasn’t enough already, this year there were more than 81,000 tech startups in the US, and 17,000 respectively in the UK. And with tech platforms constantly leap-frogging each other in terms of capability, there’s a growing need for model-agnostic tools.
In this convoluted space, the real superpower is: “connecting business data to generative AI,” says Steiner, with context becoming a company’s most valuable asset. “That’s why demand will rise for model-agnostic data orchestration and tools that let brands own and control their context. It’s exactly why we started building The Machine at Stagwell.”
The Machine is an enterprise-level, AI-driven orchestration that unifies people, tools, and data across the marketing organization – reimagining how the entire discipline operates. “We’re extremely excited about the potential of the technology that the Code and Theory team is pioneering and which will transform the way that marketers blend data and creativity to create better campaigns,” says Hong.
The last thing to look out for is the technology that allows high levels of customization and flexibility, adds Guttenberg. “We’re getting a lot from our Adobe partnership,” he says. “Their creation tools are working together to really change the velocity at which we’re going to be able to create content. We’re going to see tangible benefits of this in 2026, and it’s going to make a lot of noise.”
link